Comparison of Seegene Anyplex II HPV28 with the PGMY-CHUV Assay for Human Papillomavirus Genotyping

The PGMY primers used in PG are degenerate to identify potentially unknown genotypes.

“The PGMY primers used in PG are degenerate to identify potentially unknown genotypes (13).”

• cites [13] sourceDisputed

  The Abstract says PGMY09/11 accommodates sequence variation through pools of distinct primers, explicitly avoiding degenerate bases. The Discussion likewise describes eliminating the need for degenerate primer synthesis. The paper describes the earlier MY09/11 primers, not PGMY, as degenerate. (LLM, conf=0.90)

  Checked by a reasoning model — may be wrong.

  Experimental — accuracy not yet measured

PG was not proficient only for HPV68a, which cannot be efficiently detected with the PGMY primers.

“PG was not proficient only for HPV68a, which cannot be efficiently detected with the PGMY primers (23).”

• cites [23] sourcePartly supported

  The Discussion reports that PGMY-based assays could not detect the HPV68a plasmid because HPV68a differs at the PGMY primer-binding site. It does not establish that the assay called “PG” was proficient for every type except HPV68a; its proficiency tables report results by assay category, not that specific claim. (LLM, conf=0.84)

  Experimental — accuracy not yet measured

The higher HPV68 detection by H28 is a known consequence of the inability of the PGMY primers to efficiently target the HPV68a subtype.

“The latter is a known consequence of the inability of the PGMY primers to efficiently target the HPV68a subtype (23).”

• cites [23] sourceSupported

It is impossible to predict whether H28 would be suitable for epidemiological studies using fixed and paraffin-embedded high-grade tumor material, since sensitivity may be affected by the size and genomic location of the amplicons.

“It therefore is impossible to predict whether this assay would be suitable for epidemiological studies using fixed and paraffin-embedded high-grade tumor material, since sensitivity may be affected by the size and genomic location of the amplicons (26).”

• cites [26] sourcePartly supported

  The passage describing the study of 361 frozen cervical-carcinoma biopsies reports different detection sensitivities for primer sets targeting different HPV regions and possible L1 deletions in 23 of 56 samples. That supports concern about genomic location, but our copy does not establish an effect of amplicon size or assess whether the assay is suitable for fixed, paraffin-embedded high-grade tumors in epidemiological studies. The indexed passage has no section or page label. (LLM, conf=0.82)

  Experimental — accuracy not yet measured

These HPV51 subtypes were likely to be suboptimally detected by H28, whose sensitivity relies on type-specific primers and probes strictly adapted to published sequences.

“They therefore were likely to be suboptimally detected by H28, whose sensitivity relies on type-specific primers and probes strictly adapted to published sequences (25).”

• cites [25] sourceCould not verify

  The Discussion considers possible probe–target mismatches in influenza B assays, not detection of HPV51 subtypes by H28. The Analytical methods section describes respiratory-virus assays. Searches for HPV51, H28, and the claimed type-specific primers and probes found no passage addressing the claim. (LLM, conf=0.78)

  Experimental — accuracy not yet measured

Genotype-specific agreement was nearly perfect (total 98.71%, kappa 0.877) considering all 26 genotypes.

“While the agreement was nearly perfect (total, 98.71%; ϭ 0.877) considering all 26 genotypes, there was a statistically significant higher number of genotype-sample combinations found by H28 that were not found by PG (n ϭ 76) than of those found by PG and not by H28 (n ϭ 28, P Ͻ 0.0001).”

There was a statistically significant higher number of genotype-sample combinations found by H28 that were not found by PG (n=76) than of those found by PG and not by H28 (n=28, P<0.0001).

“While the agreement was nearly perfect (total, 98.71%; ϭ 0.877) considering all 26 genotypes, there was a statistically significant higher number of genotype-sample combinations found by H28 that were not found by PG (n ϭ 76) than of those found by PG and not by H28 (n ϭ 28, P Ͻ 0.0001).”

The higher number of H28-only genotype-sample combinations was contributed mainly by low-risk viruses (total L, 7 versus 38, P<0.0001, agreement 98.18%, kappa 0.826) in comparison with high-risk viruses (total H, 21 versus 38, P=0.0372, agreement 98.94%, kappa 0.899).

“This was contributed mainly by the low-risk viruses (total L, 7 versus 38, P Ͻ 0.0001, agreement ϭ 98.18%, ϭ 0.826) in comparison with the high-risk viruses (total H, 21 versus 38, P ϭ 0.0372, agreement ϭ 98.94%, ϭ 0.899).”

H28 was proficient for genotyping the 17 HPV types and subtypes represented in the 2011 WHO panel, including those at the lowest concentration of 5 to 50 viral genome equivalents per reaction.

“H28 therefore was proficient for the genotyping of the 17 HPV types and subtypes represented in this panel, including those at the lowest concentration of 5 to 50 viral genome equivalents per reaction.”

Among the high-risk genotypes, H28 detected significantly fewer cases of HPV51, possibly in a locally prevalent and subtype-dependent manner.

“Among the high-risk genotypes, H28 detected significantly fewer cases of HPV51, possibly in a locally prevalent and subtype-dependent manner, and more cases of HPV68.”

H28 detected more cases of HPV68 than PG.

“Among the high-risk genotypes, H28 detected significantly fewer cases of HPV51, possibly in a locally prevalent and subtype-dependent manner, and more cases of HPV68.”

The proportion of concordant genotype-sample combinations from the high-risk group (280/339) was significantly higher (P=0.0065) than from the low-risk group (114/159).

“Consistent with this observation, the proportion of concordant genotype-sample combinations from the high-risk group (280/339) was significantly higher (P ϭ 0.0065, by Fisher's exact test) than from the low-risk group (114/ 159).”

The majority of genotypes exhibited agreement values above 99%, except HPV40, -42, -51, -53, -54, -56, -58, -66, and -68.

“Individually, the majority of genotypes exhibited agreement values above 99%, except HPV40, -42, -51, -53, -54, -56, -58, -66, and -68.”

At the time of manuscript submission, PGv2 was not inferior to PG for HPV genotyping (n=315 samples, of which 7 contained HPV68a).

“At the time of manuscript submission, PGv2 was not inferior to PG for HPV genotyping (n ϭ 315 samples, of which 7 contained HPV68a).”

These HPV51 subtypes were not represented in nucleic acid databases.

“These HPV51 subtypes were not represented in nucleic acid databases.”

The lack of sensitivity of PGMY-based assays toward HPV68a is due to three mismatches within the PGMY09-L primer, as determined by DNA sequencing of a rare HPV68a PG-positive case from the authors' biobank.

“At the extreme, the lack of sensitivity of PGMY-based assays toward HPV68a is due to three mismatches within the PGMY09-L primer, as determined by DNA sequencing of a rare HPV68a PG-positive case from our biobank.”

A trend for discordance was associated with lower viral loads for HPV40, -42, and 54.

“A trend for discordance was associated with lower viral loads for HPV40, -42, and 54 (Table 4).”

There was no systematic association of two genotypes during the course of this study with either assay, even with high viral loads.

“There was no systematic association of two genotypes during the course of this study with either assay, even with high viral loads as judged by the H28 semiquantitative report and by the PG PCR/ RBH signal strengths (not shown).”

PGv2 was fully proficient with the 2013 WHO quality control panel, including HPV68a.

“PGv2 was, in addition, fully proficient with the 2013 WHO quality control panel, including HPV68a (not shown).”

Phylogenetic analysis indicated that 3 of the 4 HPV51 discordants clustered as a distinct subtype among the 10 HPV51 cases.

“Phylogenetic analysis indicated that 3 of the 4 discordants clustered as a distinct subtype among the 10 HPV51 cases (data not shown).”

Sequencing analysis of the 4 discordant single infections and of the 6 concordant single infections confirmed the HPV51 genotype.

“Sequencing analysis of the 4 discordant single infections and of the 6 concordant single infections confirmed this genotype (Table 3).”

PG was less sensitive in 3 samples involving HPV18 (5 genome equivalents, 2011 panel), HPV31 (50 genome equivalents, 2011 panel), and HPV56 (50 genome equivalents, 2013 panel).

“Although it was proficient with both panels, PG was in comparison less sensitive in 3 samples involving HPV18 (5 genome equivalents, 2011 panel), HPV31 (50 genome equivalents, 2011 panel), and HPV56 (50 genome equivalents, 2013 panel).”

H28 and PG were fully concordant at the genotypic level on 228 (73.8%) out of 309 samples: 27 HPV negative and 201 HPV positive.

“H28 and PG were fully concordant at the genotypic level on 228 (73.8%) out of 309 samples: 27 HPV negative and 201 HPV positive.”

The HPV68a single infection identified with PGv2 could be confirmed by DNA sequencing.

“The HPV68a single infection identified with PGv2 could be confirmed by DNA sequencing (Table 3).”

Among the low-risk genotypes, HPV42 was the only one displaying a significant association between discordance and the sample's multiple infection status.

“Among the lowrisk genotypes, HPV42 was the only one displaying a significant association between discordance and the sample's multiple infection status (Table 3).”

For HPV51, all discordants exhibited medium (n=6) to high (n=4) viral loads.

“For HPV51, all discordants exhibited medium (n ϭ 6) to high (n ϭ 4) viral loads (Table 4).”

Sample R50 was recorded as weakly positive by PG with stochastic amplification.

“This sample (R50) was recorded as weakly positive by PG with stochastic amplification.”

Published multiplex real-time PCR assays for HPV genotyping need at least 6 different PCR wells per sample to overcome limitations of present chemistries and equipment, which distinguish at best 3 to 5 targets per well.

“Thus, published multiplex real-time PCR assays for HPV genotyping need at least 6 different PCR wells per sample to overcome the limitations of present chemistries and equipment, which distinguish at best 3 to 5 targets per well (21).”

• cites [21] sourceNot checked — no full text of the source

Based on these results, the H28-positive/PG-negative results were indeed true positives.

“Based on these results, we concluded that the H28-positive/PG-negative results were indeed true positives.”

Enrichment of samples from patients known to have been positive at least once by PG introduced a selection bias that may have lessened the difference in sensitivity between H28 and PG.

“Enrichment of samples from patients known to have been positive at least once by PG introduced a selection bias that may have lessened the difference in sensitivity between H28 and PG.”

The trends for discordance with low viral loads and with multiple-infection status between PG and H28 regarding HPV40, -42, and -54 are in line with PGMY degeneracy.

“The trends for discordance with low viral loads and with multiple-infection status between PG and H28 regarding HPV40, -42, and -54 are in line with PGMY degeneracy.”

H28 appears very efficient for epidemiological studies using fresh cervical smears.

“H28 thus appears very efficient for epidemiological studies using fresh cervical smears.”

This observation allowed designing the HPV68a-specific RSMY09-L primer used in PGv2.

“This observation allowed designing the HPV68a-specific RSMY09-L primer used in PGv2.”

Overall, H28 performed very similarly to PG.

“Overall, H28 performed very similarly to PG.”

Like PG, H28 is too sensitive for clinical purposes, if not used as an adjunct to a clinically validated assay.

“Like PG, it is, however, too sensitive for clinical purposes, if not used as an adjunct to a clinically validated assay.”